Block Open-Sources Berd AI Agent Workspace with Local-First Storage Under Apache 2.0 License

Block Releases Berd as Open-Source AI Agent Workspace

Block, the technology company founded by former Twitter CEO Jack Dorsey, has open-sourced Berd—a desktop application originally built to give employees a unified workspace for working with AI agents across different models, tools, and projects1

1

2

2

. Available now on GitHub under the permissive Apache 2.0 license, Berd can be used, modified, and redistributed freely, including for commercial purposes. The repository reached version 0.6.2 on Aug. 18, its seventh public release, and lists 91 contributors1

1

. Free downloadable builds are available for macOS, Windows, and Linux2

2

.

Solving Internal Workflow Fragmentation

Berd emerged from a practical challenge inside Block. Employees were already working with capable agents—Block's own Goose, Anthropic's Claude Code, and OpenAI's Codex—but the experience had become fragmented. "We had capable agents through goose, Claude Code, and Codex, but working across them meant navigating different interfaces, configuration systems, and ways of managing context," Lucinda Bell, Technology Communications at Block, explained1

1

. Berd provides Block teams with one consistent desktop application across models and harnesses, eliminating the need to reconstruct context for each task2

2

.

Desktop-First Architecture for Local Work

"Berd is desktop-first because much of its value comes from working directly with projects, local files, tools, repositories, and agents running on or connected to the user's computer," Brad Axen, Head of AI Capabilities at Block, told VentureBeat1

1

. Unlike browser-based workspaces, this locally installed graphical desktop application manages chats, files, agents, models within persistent projects where users can attach files or folders, configure AI providers, manage skills and extensions, review session history, and build automations2

2

. The design emphasizes visible operational state—users can always tell which project, files, agent, model, provider, and session state are actively shaping a conversation1

1

.

Local-First Data Storage and Privacy Controls

Berd stores conversation history on the user's device in the local Goose session database, while credentials are kept in the operating system keychain by default2

2

. Axen clarified that local-first data storage does not mean no data leaves the device, because prompts and selected context are sent to whichever model provider the user configures2

2

. Telemetry is disabled by default in official Berd distributions. If users opt in, Block collects only a random installation identifier and limited predefined usage events, explicitly excluding prompts, messages, files, source code, credentials, names, email addresses, and IP addresses from the telemetry payload2

2

.

Visual Agent Identities Through Gloopies

Berd takes an unusual approach to differentiating configured agents by giving them distinct roles, instructions, skills, tools, and recognizable visual identities. Block created animated characters called Gloopies, with predefined personas including Berdy, Pushback, Choosey, Copycat, Tinker, and Wildcard1

1

2

2

. Each persona is built around a different style of agentic work: Pushback plays devil's advocate on drafts, Choosey helps narrow decisions, and Copycat learns to write in the user's style1

1

. Block stated in a company blog post: "The avatars make the agent recognizable. Its role, skills, and tools make it useful"2

2

.

Technical Architecture and Integration

Berd is built with Tauri 2 and React 19 and connects to Block's Goose agent framework through the Agent Client Protocol2

2

. Goose runs alongside Berd as a bundled sidecar process and handles the agent loop, while Berd manages projects, sessions, context, agents, and configuration2

2

. Axen emphasized that Berd is meant to sit above agent harnesses such as Goose, Claude Code, and Codex rather than replace them. In the public release, users choose their own providers and harnesses instead of using a Block-selected model stack2

2

.

Future Direction: Transition to Buzz Collaboration Platform

Block's focus is shifting toward Buzz, its open-source collaboration platform for humans and AI agents. The company plans to bring Berd's single-user features into that product. "Going forward, our focus is Buzz, and we plan to bring the best parts of Berd's single-player experience into it," Axen said2

2

. Block has not announced a subscription price for Berd, and the software remains free and open-source, though users still pay charges from model providers they choose2

2

. Axen indicated that enterprise-focused commercial layers around enterprise deployment may come later, adding: "Supporting organizations that want to deploy agents at work is an area of interest for us"2

2

. The repository does not accept outside pull requests, and external code contributions are automatically closed, while users can inspect, build, and fork the software as the upstream project remains under the control of a small internal team2

2

.